A Troxler neutron soil moisture meter was 

 used to measure the moisture content at pre- 

 determined depths in each tank under condi- 

 tions of changing water levels. Thirteen water 

 levels, measured in inches from the surface of 

 the sand, were selected for this study (fig. 5). A 

 one-minute counting period (Merriam 1962) 

 was used at each of these 13 water levels. The 

 center of measurement for the probe was as- 

 sumed to be the center of the detector tube. 

 After a set of 13 one-minute counts, the water 

 level was lowered or raised to the next water 

 level and the neutron measurements were re- 

 peated at all 13 water levels. Each set of read- 

 ings was bounded by a set of 5 one-minute 

 standard counts. The water which drained from 

 the tank when the water level was lowered to 

 the next stage was collected, measured, and 

 compared with the weight of the tank before 

 and after the change in water level. When the 

 neutron measurements were completed on one 

 tank (i.e., from a saturated to a drained condi- 

 tion), the process was repeated on the second 

 tank. After all neutron measurements were 

 completed, a series of bulk density samples 

 were taken; the tanks were emptied and refilled 

 with the fine sand. 



Three sets of bulk density measurements 

 were made in each of the two tanks: (1) small 

 (142 cc.) gravimetric samples were taken 

 through the side of the tank; (2) measurements 

 were made with a Troxler depth -density gage; 

 and (3) large (6,300 cc.) gravimetric samples 

 were taken adjacent to each void when the 

 tanks were emptied. 



The small sampler consisted of a length of 

 stainless steel conduit having an inside diameter 

 of 1.05 inches; it was machined on the outside 

 to a 1.12-inch diameter and had a flange acting 

 as a stop to allow the sampler to penetrate 10 

 inches into the sand. After the tank had drained 

 for 24 hours, a plug was removed from a sam- 

 pling hole, and the small sampler was inserted. 

 The 142-cc. sample was withdravm, emptied 

 into a can, weighed, and placed in an oven for 

 drying and reweighing. As soon as the sampler 

 was emptied, the sand was replaced by a sample 

 drawn from a container of damp sand filled to 

 the same bulk density from the overhead 

 hopper. The replacement was inserted into the 

 tank and held in place by a wooden plunger 



Figure 5. — Darker portion of tank on right is caused 

 by rising water. Drawing on side of tank shows size 

 and location of the small void and the large void with- 

 in the tank. Numbers are in inches, measured from 

 the surface of the medium. 



while the sampler was vnthdrawn and the rub- 

 ber plug was replaced. 



The second set of samples was taken by low- 

 ering the Troxler depth-density gage into the 

 access tube and measuring at each of the 13 

 levels. 



The large gravimetric samples were taken at 

 four levels in each tank. The first sample was 

 taken after 4 inches of sand had been re- 

 moved from the tank. The sampler was forced 

 into the sand, the surrounding seind was re- 

 moved, a piece of sheet metal was inserted 

 under the sample, and the sample was removed 

 from the tank; the sample was weighed, dried, 

 and then reweighed. Another sample was taken 

 between the small void and the wall of the tank; 

 the center of the sample was on a level with the 

 center of the void. A third sample was taken on 

 a level corresponding to the center of the tank, 

 and a fourth sample was taken opposite the 

 large void. 



5 



